Vibrational spectroscopy of the ammoniated ammonium ions NH4+

Patrick J. Kelleher , Christopher J. Johnson , Joseph A. Fournier , Mark A. .... Y. T. Lee and B. H. Mahan, “Mobilities of Cesium and Rubidium Ion i...
4 downloads 0 Views 2MB Size
2182

J . Phys. Chem. 1991, 95, 2182-2195

Vibrational Spectroscopy of the Ammonlated Ammonlum Ions NH,+(NH,), ( n = 1-10) J. M. Price, M. W. Crofton, and Y.T. Lee* Materials and Chemical Sciences Division, Lawrence Berkeley Laboratory and Department of Chemistry, University of California, Berkeley, California 94720 (Received: May 17, 1990)

The gas-phase vibration-internal rotation spectra of mass-selected ammoniated ammonium ions, NH4+(NH3), (for n = 1-10), have been observed from 2600 to 4000 cm-I. The spectra show vibrational features that have been assigned to modes involving both the ion core species, NH4+,and the first shell NH3 solvent molecules. Nearly free internal rotation of the solvent molecules about their local C3 axes in the first solvation shell has been observed in the smaller clusters ( n = 1-6). For the largest clusters studied ( n = 7-10) the spectra converge, with little difference between clusters differing by one solvent molecule. For these clusters, the spectrum in the 3200-3500 cm-' region is quite similar to that of liquid ammonia, and the entire region of 2600-3500 cm-' alsobears considerable resemblance to the spectra of ammonium salts dissolved in liquid ammonia under some chemical conditions. This indicates the onset of a liquidlike environment for the ion core and first shell solvent molecules in clusters as small as NH4+(NH3),.

Introduction

Early spectroscopic studies of the ammonium ion were confined to the condensed phases and employed direct IR absorption measurements of either crystals of ammonium salts or ammonium salts dissolved in s~lution.l-~ These measurements revealed spectral features that were assigned to vibrational modes of the perturbed NH4+ ion (Td symmetry). Of these, the strongest in the 2000-4000-cm-' region were found to involve the triply degenerate antisymmetric stretching motion u3 and the first overtone of the doubly degenerate bending motion 2u4 (see Figure 1). Gas-phase studies which followed centered for the most part on the thermodynamics and kinetics of clustering reactions at relatively high pressures. Many investigators contributed in the effort to determine the AH0 for formation of the solvated ammonium ion N H4+(NH3),,.*-I0 Kebarle and co-workers showed that in the larger clusters there was a discontinuity in the AH0 for the n = 4 n = 5 clustering step. This result was consistent with the idea that there were four solvent molecules involved in the first solvation shell of the cluster, with the solvent ammonias probably oriented along the N-H bonds of the ammonium ion due to ion-dipole interaction. A summary of these results appears in Table I, which lists the AH0 of formation for each clustering step in the NH4+(NH3),+ NH3 s NH4+(NH3),,+l reaction series. A fairly extensive body of theoretical work also exists, investigating the structure and thermodynamics of these systems. Pullman and Armbruster examined the thermodynamics of the stepwise addition of H 2 0 or NH, to NH4+for one to five solvent molecules and managed to reproduce the qualitative features of the experimental data." The experimental change in preferential addition from NH, at low levels of solvation to H 2 0 at higher levels of solvation was reproduced. Errors in the energies calculated from the STO 3G basis set were found to decrease with increasing number of solvent molecules. A more recent series of calculations on the thermodynamics of the NH4+(NH3),( n = 1-5) ions as well as their detailed structures have been made by Hirao et al.I2 In this work, ab initio

-

( I ) Wagner, E. L.; Horning, D. F. J . Chem. Phys. 1949, 18, 296. (2) Mathieu. J. P.; Poulet, H. Spectrochim. Acto A 1960, 16, 696. (3) Corset, J.; Huong, P. V.;Lascombe, J. Spectrochim. Acto A 1968, 24,

2045. (4) Hogg, A. M.; Kebarle, P. J . Chem. Phys. 1965, 43,449. ( 5 ) Hogg, A. M.; Haynes, R. M.; Kebarle, P. J . Am. Chem. SOC.1966, 88, 28. ( 6 ) Searles, S. K.; Kebarle, P. J . Phys. Chem. 1968, 72, 742. (7) Payzant, J. D.; Cunningham. A. J.; Kebarle, P. Con. J . Chem. 1973, 51, 3242. (8) Tang, 1. N.; Castleman Jr., A. W . J . Chem. Phys. 1975, 62, 4576. (9) Arshadi, M. R.; Futrell, J. H. J . Phys. Chem. 1974, 78, 1482. (IO) Ceyer, S . T.; Tiedemann, P. W.; Mahan, 8 . H.; Lee, Y. T. J . Chem. Phys. 1979, 70 ( I ) , 14. (11) Pulman, A.; Armbuster, A. M. Chem. Phys. Lett. 1975, 36, 558.

0022-3654/91/2095-2182$02.50/0

TABLE I: Thermodynamic Data (in kcal/mol) for Gas-Phase Clustering ReactioniNH4+(NH3)" + NH3 2 NH4+(NH3),+, n. n 1 AC",,," AH"W,~~ AHOm' AH0nd 0, 1 -17.5 -27 -25.4 -21.5 -13.8 1, 2 -17 -17.3 -16.2 -6.4 2, 3 -6.4 -16.5 -17.3 -13.5 3, 4 -3.8 -14.5 -14.2 -11.7 5 -0.2 -7.5 -11.8 -7.0 4,

+

"Searles, S.K.; Kebarle, P. J . Phys. Chem. 1968, 72, 742. bTang, I. N.; Castleman, Jr., A. W. J . Chem. Phys. 1972, 62, 4576. 'Arshadi, M. R.; Futrell, J. H. J . Phys. Chem. 1974, 78, 1482. dCeyer, S.T.; Tiedemann, P. W.; Mahan, B. H.; Lee, Y . T. J. Chem. Phys. 1979, 70, 14. TABLE 11: Molecular Constantsa for NH4+ ( T d Symmetry) Y , : 3270 25b u2: 1669" v j : 3343.28/ u4: 1447.229 A = B = C: 5.9293 f 0.0002f 0.0606

*

c3:

NHAt and NH1

NH3 (C3"Symmetry) u,: 3336.21' ~ 2 : 950'

u3:

3444e

2 u f i 3216.71' 2u': 3240.4S B = C: 9.941," A : 6.30gh 0.06h u4: 1627:

c3:

"Units are cm-I for all constants except 13, which is dimensionless. * Botschwina, P. J . Chem. Phys. 1987, 87, 1453; scaled ab initio value. 'Benedict, W. S.; Plyler, E. K.; Tidwell, E. D. J . Chem. Phys. 1960, 32, 32. "DeFrees, D. J.; McLean, A. D. J . Chem. Phys. 1985,82, 333; scaled ab initio value. 'Shimanouchi, T. Tables of Molecular Vibrarional Frequencies; Natl. Stand. Ref. Data Ser., Natl. Bur. Stand. N o 39; US. Government Printing Office: Washington, DC, 1972. 'Reference 14. fpolak, M.; Gruebele, M.; DeKock, B. W.; Saykally, R. J. Mol. Phys. 1989, 66, 1193. See ref 38, p 431. molecular orbital calculations determined the structure and binding energies of the series in reasonable agreement with high-pressure equilibrium measurements. Values for the solvation energies obtained from this work, using 3-21G, 3-21G*, and 6-31G** basis sets, yielded consistently higher values than experimentally obtained. Gas-phase spectroscopic measurements for the ammonium ion and the ammoniated ammonium clusters have only recently been carried out. The development of velocity modulation spectroscopy has yielded a great deal of information for a wide range of molecular ions, including the NH4+ ion.I3 The u3 band of the ammonium ion has been investigated at high resolution by both Crofton and Oka14 and Schafer and SaykallyIs using this tech(12) Hirao, K.; Fugikawa, T.; Konishi, H.; Yamabe, S. Chem. Phys. Lett.

1984, 104, 184. (13) Gudeman, C. S.; Saykally, R. J. Annu. Reo. Phys. Chem. 1984,35,

387.

0 1991 American Chemical Society

Vibrational Spectroscopy of NH4+(NH3), (n = 1-10)

The Journal of Physical Chemistry, Vol. 95, No. 6, 1991 2183

Grid

Corona Discharge Tip

t I

All

I

1

All

I

I

I

lm I

I

b

+ Figure 1. Normal vibrations of tetrahedrally symmetric NH4+. ug and u, are 3-fold degenerate, while u , and u2 are 1- and 2-fold degenerate, respectively. Arrows in the figure indicate the direction of motion but are not to scale with the amplitudes. (See ref 38.)

nique. From these data on accurate equilibrium geometry has be& obtained.I4 Table I1 lists the measured rotational constants and equilibrium geometry parameters for this ion and the NH3 molecule. The only measurements of the spectra of the ammoniated ammonium ions ( n = 1-4) have been limited to the direct absorption measurements performed by Schwarz in 1980 in a pulsed radiolysis of a gas cell containing ammonia in a helium carrier.I6 Schwarz's measurements made use of the known equilibrium constants for the clustering reactions to deconvolute the absorption signal obtained at various pressures into those of individual ammoniated ammonium ions. These pioneering measurements, discussed later in the context of the results of the present study, suffered from limited spectral resolution (40-cm-I full width at half-maximum (fwhm)), relatively high vibrational and rotational temperatures (>300 K), and some ambiguity regarding the distribution of cluster sizes present in the sample. Nonetheless, it was possible for Schwarz to observe the vibrational bands u3 and 2u4 of the ammonium ion core for the n = 4 cluster and to show that the n = 3 spectrum was consistent with C3,symmetry. Schwarz was able to prove by isotope substitution that the structure of the n = 4 ion cluster is tetrahedral. No transitions were observed that could unambiguously be assigned to the solvent molecules surrounding the ammonium ion. Recently, we have studied the infrared vibrational predissociation spectroscopy of the tetraammoniated ammonium ion NH4+(NH3)4using an arrangement of tandem mass spectrometers, a radio-frequency (rf) ion trap, and a tunable infrared laser.'' In addition to simply vibrational transitions, we have observed spectral features that are due to nearly free internal rotation of the ammonia subgroups in NH4+(NH&, the first observation of such a phenomenon in an ionic cluster. This work also served to verify Schwarz's spectrum determined by deconvolution of the time-dependent infrared absorbance based on equilibrium and rate constants and his assignment of the v3 and 2u4 bands in the spectrum. The comparison with our spectra further indicates that the spectra we derive from vibrational predissociation are similar to those obtained by direct absorption for this system. In the present work we give a full account of our systematic investigation of a series of infrared vibrational predissociation (IRVPD) spectra of the ammoniated ammonium ions and Dresent the absorDtion spectra for NH4+(NH3), ( n = 1-10) ove'r the frequency ;ange 2600-3500 cm-I. It will be shown that correlation of spectral features with cluster size can provide information about solvation shell structure, the geometry of the clusters, and the onset of a condensed-phase

I

Source Body Figure 2. Schematic diagram of the ion source, which consists of a corona discharge and supersonic expansion. Corona Discharge lon Source

Laser

-A

--

.- .......................P ..........................OOOO ........q a

oooo o r l a s e r

II

Octopde lon Trap

lon Detector

Figure 3. Schematic diagram of the experimental apparatus.

environment around the ion core. With sufficient laser resolution, rotational information about the cluster should also be obtainable with this new experimental technique. Experimental Details

The experimental apparatus used in this work has been described previousIy.l8-*' Briefly, the ammoniated ammonium cluster ions were produced from a high-pressure corona discharge source and subsequent supersonic expansion. Schematic diagrams of the source and the experimental apparatus are shown in Figures 2 and 3. The corona discharge was maintained in 100-200 Torr of gas (Matheson ultrahigh purity (UHP) neon typically seeded with 10% U H P H, and about 1% NH,) flowing past a 1.2-kV potential from the discharge tip of the needle to the source body maintained at approximately 350 V above ground. The discharge current under these conditions ranged from 40 to 200 p A . The source could be cooled from outside the machine by contact with a liquefied gas reservoir and was usually maintained at approximately -30 "C.After passing through the discharge region, the ionized gas undergoes collisionally induced vibrational relaxation in the 1 .O-mm-long by 3.0-mm-diameter drift region before expanding through the 70-pm expansion nozzle. It is estimated that an ion undergoes some lo5 collisions in this region before leaving the source, resulting in the ion relaxing to a vibrational temperature near that of the source body by the time it reaches the nozzle. Most of the clustering of the neutral gas around the ionic species and rotational cooling takes place during the supersonic expansion between the nozzle and the skimmer. Typical pressures in the ~~~~~

(14) Crofton, M. W.; Oka, T. J . Chem. Phys. 1983, 79, 3157; 1987,86, 5983. (15) Schafer, E.; Saykally, R. J. J . Chem. Phys. 1984. 80 (9). 3969. (16) Schwarz, H. A. J . Chem. Phys. 1980, 72, 284. (17) Price, J. M.;Crofton, M. W.;Lee, Y . T. J . Chem. Phys. 1989, 91, 2749.

(18) Bustamente, S. W. Ph.D. Thesis, University of California at Berkeley, 1983. (19) Bustamente, S. W.; Okumura, M.; Gerlich, D.; Kwok, H. S.; Carlson, L. R.; Lee, Y. T. J . Chem. Phys. 1987,86, 508.

(20) Okumura, M. Ph.D. Thesis, University of California at Berkeley,

1986. (21) Yeh, L. 1 . C Ph.D. Thesis, University of California at Berkeley, 1988.

2184 The Journal of Physical Chemistry, Vol. 95, No. 6, 1991

source chamber are 4 X IO" to 2 X IO4 Torr while running the experiment. To prevent internal excitation and dissociation of the ionic clusters through collisions with the background gas in the expansion, the potential of the skimmer was maintained within 5 V of that of the source body. A shielding grid surrounding the expansion region maintained at 350 V prevented stray fields from affecting the ion trajectories in this region. After the skimmer, the ion beam enters a second differential pumping region containing collimating and focusing optics. The pressure in this second region is typically an order of magnitude lower than that of the source region. The beam is then directed into a third differentially pumped region maintained at Torr and through the entrance slits of a 60' sector magnet mass analyzer (resolution = M / A M = 200). To aid in transmission of the ion beam and to enhance mass resolution, a set of quadrupole lens pairs is placed before and after the nagn net.*^,^^ The mass-selected beam passes into the final differentially pumped region maintained at IO4 Torr. Here, the beam is bent 90° in a dc quadrupole field, decelerated to less than 0.5 eV, and focused into a rf octopole ion trap through an entrance aperture lens. Ions are typically allowed into the trap for 1 ms before the potential of the entrance lens is raised, and the ions are confined inside the octopole. The confined, mass-selected clusters are then vibrationally excited by a pulsed, tunable infrared laser. (Quanta Ray, IR WEX 8-ns pulse, 0.3-1.2-cm-' resolution, 0.7-cm-' absolute frequency accuracy, 0.2-1 .O mJ/pulse in the 2600-4000-cm-' region). The density of ionic clusters in the ion trap is not high enough to allow the measurement of photon absorptions directly. In order to detect the absorption of an IR photon by an ionic cluster, IR multiphoton dissociation processes were used to exclusively dissociate the vibrationally excited ionic clusters. Depending on the density of states and the dissociation energy of the species under study, one of two excitation schemes described below was employed. Because of the large binding energies of the n = 1 and 2 ammoniated ammonium ions, absorption of sufficient energy from the tunable infrared (IR) laser to cause predissociation is not a facile process. I n these cases, a line tunable C 0 2 laser is used (MPB Technologies Inc., 8 W CW on R(24) of the 00°1-0200 transition) to drive the clusters excited by the tunable IR laser over the dissociation limit through the absorption of multiple C 0 2 laser photons. This process hinges on the initial absorption of the tunable IR photon, as the density of states 1000 cm-' above the u = 0 level is so low that multiphoton dissociation of the ionic clusters by the CW CO, laser alone is very unlikely even with tens of milliseconds of irradiation time. In the studies of the n = 1 cluster, no vibrational predissociation signal was observed if the pulsed laser was not present. Further discussion of this technique may be found elsewhere.*' For the larger clusters, n = 3-10, the energy required to predissociate one solvent molecule is exceeded by absorbing two photons in the usual tuning range of 2.5-3.8 Hm. At an energy in the 2600-4000-cm-' region, all of the ionic clusters ( n = 1-10) are literally in the "quasi-continuum" region24after the absorption of one photon, and an additional IR photon from the tunable laser may be absorbed to induce vibrational predissociation. For n = I and 2, however, fragments were not detectable with the tunable laser only, presumably because the fluence was not sufficient for the absorption of more than two photons. For these ions, use of the C 0 2 laser with an irradiation time of at least 10 ms at >IO W/cm2 was required to achieve a measurable degree of dissociation of vibrationally excited clusters. In those instances where spectra of a given ion were obtained by both one- and two-color ~

~~~~

(22) Enge, H. A. Reo. Sei. Instrum. 1959, 30, 248. Giese, C. F. Ibid. 1959, 30, 260. Lu, C.-S.; Carr, H. E. Ibid. 1962, 33, 823. (23) Taya, S.; Kanomatata, 1.; Hirose. H.; Noda, T.; Matsuda, H. Int. J . Mass Specfrom. Ion Phys. 1978, 26, 231. (24) Demtrbder, W. Laser Specfroscopy: Basic Concepts and Instrumentation; Springer-Verlag Press: New York, 1982. Letokhov, V. S. Nonlinear Lmer Chemistry: Multiple-Photon Excitation; Springer-Verlag Press: New York, 1983.

Price et al. I

/A.

I

I

1

1

1

I

I

2

3

4

I

I

I

I

I

I

I

I

8 Number Solvent NHjs 5

6

7

I

I

I

I

I

I

9

1011

Figure 4. Mass spectra showing the NH4+(NH3),cluster distribution as a function of ion source conditions. For (a), the NH,/H,/Ne mixture flowed through a liquid nitrogen trap to reduce the NH3 concentration. The source backing pressure was raised to obtain (b), and the trap was bypassed to obtain (c). The NH3/H2/Ne mixture was =1:10:100.

schemes, they were found to be the same within experimental error, indicating that spectral features in the multiphoton dissociation spectra normally reflect the cross section for absorption of the first photon rather well.21 If the clusters absorb sufficient energy through one of the schemes described above, the loss of one or more solvent molecules from the parent cluster may occur. The potential on the exit aperture is lowered 1 ms after the laser pulse, extracting cluster ions of all masses from the trap. These ions are filtered by a second mass analyzer, a quadrupole mass spectrometer tuned to pass only the daughter ions smaller than the parent cluster by one solvent molecule. Daughter ions are counted by means of a Daly ion detector25 for each laser shot. Background daughter ion counts resulting from the decay of metastable parent ions in the rf ion trap are monitored in a separate data cycle with the laser off at each wavelength and subtracted from the laser on signal. Typical background count rates usually amount to no more than 1% of the signal at the stronger absorption maxima. The spontaneous decay of metastable parent ions to daughter ions in the trap was as high as 5% for the largest ions and as low as for NH4+(NH3). Laser power is monitored at each data point, and spectra are normalized for the tunable infrared power by using a simple linear power dependence. For a typical experiment, signals were averaged for about 400 laser shots at each wavelength in the 2600-3200-cm-' region and longer and about 1000 laser shots at each wavelength in the 3200-4000-cm-' region where signals were typically much weaker. This extra averaging is reflected in the better signal-to-noise ratios in the higher frequency region for some of the traces shown in Figure 9. Relative intensities between the two frequency regions should be correct.

Results and Discussion A. Mass Spectra. The distribution of ionic clusters produced by the corona discharge source could be shifted by altering the temperature of the source body, the backing pressure of the supersonic expansion, or the percentage of NH3 in the H2 carrier gas. Little dependence was observed between the peak of the cluster ion distribution and the magnitude of the current used in the discharge for a given value of the above parameters, indicating that the ions were at a reasonable equilibrium with the temperature of the source body before expansion through the nozzle. The distribution of cluster sizes produced by the source could be easily monitored by sweeping the field of the 60' sector magnet and counting the ions that arrive at the Daly detector with the second mass filter set to transmit all ions. Two such mass spectra (25) Daly, R. N. Rev. Sci. Instrum. 1960, 31, 264.

Vibrational Spectroscopy of NH4+(NH3),( n = 1-10) are shown in Figure 4. Spectrum a was taken with a low concentration of ammonia in the H2 carrier gas (